The display of polypeptides on the surface of genetic packages represents a powerful methodology for carrying out molecule evolution in the laboratory. The ability to construct libraries of enormous molecular diversity and to select for molecules with desired properties has made this technology applicable to a wide range of problems. The origins of phage display date to the mid-1980s when George Smith first expressed an exogenous segment of a protein on the surface of bacteriophage M13 virus particles by fusing the exogenous sequence to a phage coat protein (Science (1985) 228: 1315-1317). Two groundbreaking concepts emerged from Smith's initial experiment. First, the experiment suggested that a vast diverse repertoire of polypeptides could be constructed in which individual phage particles display unique polypeptides. Second, the experiment confirmed a direct physical link between phenotype and genotype. That is, the phage displaying the desired polypeptide also harbors the DNA encoding the polypeptide, which can be readily isolated for subsequent analyses. McCafferty and Ladner extended these concepts to screen repertoire of polypeptides such as single-chain antibodies displayed on the surface of phage particles (U.S. Pat. Nos. 5,969,108 and 5,837,500). Since then, phage display has become a popular technique for protein engineering.
A range of display systems have been developed based on George Smith's findings. These systems can be broadly classified into two categories. The first generation system is a one-vector system. The vector in this system contains the entire phage genome, insert therein an exogenous sequence in-frame with a coat protein gene. Because the resulting phage particles carry the entire phage genomes, they are relatively unstable and less infectious. The second generation system, commonly referred to as the phagemid system, has two components: (1) a phagemid vector carrying the exogenous sequence fused to phage coat protein, and a phage-derived origin of replication to allow packaging the phagemid into a phage particle; and (2) a helper phage vector carrying all other sequences required for phage packaging. The helper vector is typically replication-defective such as M13KO7 helper vector manufactured by Amersham Pharmacia Biotech and its derivative VCSM13 that is produced by Stratagen. Upon superinfection of a bacterial cell with the helper phages, newly packaged phages carrying the phagemid vector and displaying the exogenous sequence are produced.
As such, the prior phagemid system requires fusion of the exogenous sequence to at least part of a phage outer-surface sequence (i.e. the coat sequence). The fusion or display sites most commonly used are within genes III and VIII of M13 bacteriophage, although genes VI, VII and IX fusions have been reported. However, these fusion systems bear a number of pronounced limitations. First and foremost, the expression of coat proteins is toxic to the host cells, thus tight regulation of the coat-fusion must be monitored. Even so, the unavoidable promoter leakage can cause loss of members of a diverse library. Maintaining the stability of a library is especially critical for generating a vast diverse repertoire of molecules (such as antigen-binding units) with a complexity of at least 109. Second, expression of certain coat proteins such as the gene III product (pIII) can render host cells resistant to infection with helper phage-required for the production of progeny phage particles. Third, the fusion format including the gene III and gene VIII phage display systems restrict the point of insertion to the 5′ end of the outer-surface sequence. The exogenous polypeptide thus must be linked to the N-terminus of the outer-surface proteins. Consequently, cDNA libraries containing fragments of coding sequences of all reading frames cannot be fully displayed by these fusion systems due to frequent disruption of reading frames by internal stop codons. Furthermore, the fusion system is unstable due to recombination between the fusion and the wildtype outer-surface protein that is typically provided by a helper vector. Finally, since the phagemid vector contains at least a portion of the outer-surface sequences, large exogenous sequence may not be efficiently expressed because of low transformation efficiency of a large vector. Transformation efficiency, however, is a critical factor for the production of libraries of high complexity.
Various modifications to the fusion phagemid system have been described. WO 91/17271 proposes construction of a phage display system in which the exogenous sequence is displayed via interaction of a “tag” and a “tag ligand.” The system contains a phage genomic vector that carries the exogenous sequence joined to a tag sequence. The same vector carries a tag ligand sequence fused in-frame with a coat protein gene. Upon infection of a host cell with the vector, it is speculated that phage particles expressing the exogenous sequences would be produced. However, the disclosure of WO 91/17271 does not provide a teaching which enables the general idea to be carried out. For example, WO 91/17271 does not demonstrate that any sequence has been displayed on the surface of phage particle via the interaction between a “tag” and a “ligand;” nor has it demonstrated that the protein, if expressed, retains biological activity. Furthermore, because the proposed system employs a phage genomic vector carrying all phage coat protein genes in the same vector, the system inevitably inherits all limitations and drawbacks as described above.
Crameri et al. devised a system to display cDNA products, in which Fos oncogene was inserted adjacent to the exogenous sequence to be displayed on a phagemid vector, and Jun oncogene was inserted adjacent to gene III on the same vector (see Crameri et al. (1993) Gene 137:69-75). These two fusion sequences were placed under the control of two separate promoters. The Crameri approach exploits the preferential interaction between fos and jun proteins: as the Fos-exogenous polypeptide is expressed and secreted into the periplasmic space, it forms a complex with pIII-Jun which is then packaged into the phage particles upon superinfection with M13KO7 helper phage. Although the exogenous sequence in this system is not directly linked to an outer-surface sequence, the constitutive expression of phage coat protein pIII under a separate promoter of the same vector still causes substantial toxicity to the host cells.
Another variant similar to the Crameri system is the “cysteine-coupled” display system described in WO 01/05950. The attachment and display of the exogenous polypeptide are mediated by the formation of disulfide bond between two cysteine residues, one of which is contained in the exogenous sequence, and the other is inserted in the outer-surface sequence. The one vector system described in WO 01/05950 is a phagemid vector carrying two separate promoter-controlled expression cassettes: one expresses the exogenous sequence, and the another expresses the coat protein pIII. The two-vector system described in WO 01/05950 contains a phagemid vector carrying an exogenous sequence, and a plasmid expressing the coat protein pIII. The two vectors are used to co-transfect E. Coli cells. Upon superinfection with the helper phages, M13KO7, the phagemid and/or the plasmid are packaged into the resulting phage particles. Although this system avoids the expression of a fusion comprising the exogenous protein linked to an outer-surface protein, the system again fails to minimize the toxicity of coat proteins to the host cells because of the constitutive expression of the coat protein pIII in either the one-vector or the two-vector system. Furthermore, the two-vector system described in WO 01/05950 inevitably produces phage particles with mispackaged vectors carrying the outer-surface sequences and not the exogenous gene upon infection of the helper phages. Mispackaging is a well-known problem associated with two-vector system. It has been shown that the pIII-supplementing plasmid vectors were mispackaged into helper phage particles (Rondot et al. (2000) Nature 19: 75-78).
Finally, the aforementioned prior phage display systems are not compatible with other display systems, such as a bacterial display system. To present the same phage-displayed exogenous sequence directly onto a bacterial cell, the exogenous sequence must first be subcloned into a bacterial display vector.
Thus, there remains a considerable need for improved compositions and methods for exogenous display on genetic packages. An ideal system would avoid the drawbacks of the previously reported systems. The present invention satisfies these needs and provides related advantages as well.